Introduction

Plant resistance can be generalized into qualitative and/or quantitative interactions between plant hosts and pathogens (Kushalappa et al. 2016). Much research has described the genetics and molecular mechanisms underlying qualitative interactions, which are predominantly mediated by resistance (R) genes that follow a gene-for-gene inheritance model (Flor 1942). While there are many examples of qualitative resistance to foliar diseases caused by biotrophic fungi (Glazebrook 2005), quantitative interactions are generally found with diseases caused by hemibiotrophic or necrotrophic fungi due to complex genetics and molecular interactions (Poland et al. 2009).

Genetic mapping for quantitative interactions often results in multiple quantitative trait loci (QTL) associated with disease resistance, making the breeding for quantitative resistance technically challenging as each of the QTL may provide a small contribution to the overall expression of resistance (St. Clair. 2010). Sudden death syndrome (SDS) is a soybean [Glycine max (L.) Merrill] disease caused by the soil-borne fungus Fusarium virguliforme, and SDS resistance is quantitatively inherited (Roy et al. 1997; Hartman et al. 2015). Genetic mapping for SDS resistance is complex and more than 80 QTL have been discovered from linkage mapping studies and some of these QTL are documented in SoyBase (Grant et al. 2010; https://soybase.org). With the recent advances in genome-wide association studies (GWAS), dozens of single nucleotide polymorphisms (SNPs) for SDS resistance have also been reported (Bao et al. 2015; Chang et al. 2016a; Wen et al. 2014; Zhang et al. 2015). The volume of information has made SDS mapping results complicated and hard to follow. Reasons for the complexity lie in the differences in phenotyping methodologies, the etiology of SDS foliar and root symptoms, the potential disease synergism between F. virguliforme and soybean cyst nematode (SCN; Heterodera glycines), the interactions among F. virguliforme, soybean genotypes, and environments, and the intricacies of the QTL nomenclature. Because of these reasons, a validation system needs to be established to ascertain which QTL are reliable. This review summarizes and organizes SDS resistance mapping literature up until 2017. It communicates the basis of SDS resistance loci in the soybean genome and provides a guideline for researchers involved with improving SDS resistance.

SDS disease resistance research: complexities

Etiology of SDS foliar symptoms and root rot

In the US, SDS is caused by a soil-borne fungal pathogen, F. virguliforme, which infects soybean roots and causes root rot throughout the growing season. F. virguliforme also produces phytotoxins (Brar et al. 2011; Chang et al. 2016b), which translocate to leaves causing SDS foliar symptoms usually observed at during later growing stages, typically around flowering through to pod fill (Hartman et al. 2015). Typical SDS foliar symptoms, including interveinal chlorosis and necrosis, are age related and time dependent (Chang et al. 2016b; Gongora-Canul and Leandro 2011; Kandel et al. 2016b; Vosberg et al. 2017). SDS occurrence and its severity depend on several factors, including soybean genotypes (Hartman et al. 1997; Kandel et al. 2016b; Mueller et al. 2002; Rupe et al. 2000; Vick et al. 2006), rotation and tillage (Rupe et al. 1997; Vick et al. 2003, 2006; Xing and Westphal. 2009), soils (Chong et al. 2005; Gongora-Canul and Leandro 2011; Scherm and Yang. 1996; Scherm et al. 1998), irrigation and flooding (Abdelsamad et al. 2017; Farias Neto et al. 2006; Scherm and Yang. 1996), fungicide and herbicide treatments (Kandel et al. 2015, 2016a, b; Sanogo et al. 2001; Vosberg et al. 2017; Weems et al. 2015), SCN incidence in fields (Brzostowski et al. 2014; Gao et al. 2006; Kandel et al. 2017; Marburger et al. 2014; Roth et al. in press; Rupe et al. 1999; Westphal et al. 2014; Xing and Westphal 2006), the soil microbiome (Srour et al. 2017), and other environmental factors (Adee et al. 2016; Rogovska et al. 2017; Rupe et al. 2000, 2013; Scherm and Yang 2010). These complex interactions may be the cause of variable heritability (ranging from 0.1 to 0.9) reported in SDS mapping literature (Anderson et al. 2015; Chang et al. 1996; Farias Neto et al. 2007; Hnetkovsky et al. 1996; Kazi et al. 2008; Luckew et al. 2017; Meksem et al. 1999; Njiti et al. 1996, 1998, 2001; Njiti and Lightfoot 2006; Wen et al. 2014), suggesting that a rigorous and appropriate phenotyping methodology needs to be carefully considered.

Phenotyping methodologies for SDS resistance

There are at least three different SDS phenotyping methodologies, including a stem cutting assay, greenhouse screening, and field screening, but phenotypic ratings are done mostly on the basis of foliar symptoms because foliar symptoms are distinct and easily observable. However, SDS resistance is composed of two parts, the foliar responses to phytotoxins and the root responses to infection (Kazi et al. 2008; Lightfoot 2015; Ortiz-Ribbing and Eastburn 2004). It has been shown that the severity of foliar symptoms has low correlation with the severity of root rot (Bao et al. 2015; Mueller et al. 2002), and transcriptome analysis of foliar responses to phytotoxins and root responses to infection was also different (Radwan et al. 2011, 2013). Accordingly, soybean leaves and roots could have an independent quantitative interaction to SDS.

Phenotyping soybean trifoliolates in the stem cutting assay specifically tests for foliar sensitivity to phytotoxins (Hartman et al. 2004; Li et al. 1999; Xiang et al. 2015), but the method obviously does not take into account root infection or root resistance. Accordingly, the stem cutting assay provides a specific phenotyping for foliar resistance, but it may or may not reflect field performance.

Phenotyping soybean in the greenhouse is often completed by mixing soils with F. virguliforme-infested sorghum or cornmeal inoculum (Farias Neto et al. 2008; Gongora-Canul et al. 2012; Scandiani et al. 2011). The method provides even distribution and equal chance of infection for each soybean genotype. The result from a greenhouse screening represents a combined performance of foliar and root resistance. While foliar rating is straightforward, root rot needs to be assessed by washing roots and rating for severity. Because root rot development can occur throughout the growth of the plant, the best evaluation timing of foliar symptoms and root rot should be considered separately.

Phenotyping soybean in fields with a history of SDS incidence may provide a closer environment to real-world farming conditions, but the heterogeneity of inoculum and soil properties in the field may cause patchy disease occurrence. To address the uniformity of SDS expression in fields, mixing seed with F. virguliforme-infested sorghum at planting is one technique to enhance field screening. Assuming inoculum supplement provides sufficient and even biotic stress and the environment is suitable for foliar symptoms occurrence, an observation of no SDS foliar symptoms or a few symptoms may indicate three possibilities: (1) the soybean genotype possesses foliar resistance although roots were infected. Foliar symptoms are not present because of leaf insensitivity to phytotoxins; (2) the soybean genotype has root resistance to pathogen infection, but the leaves may be sensitive to phytotoxins if the genotype is tested with the stem cutting assay; or (3) the soybean genotype carries both foliar and root resistances. The three foliar phenotype scenarios underlie the insufficiency of relying only on rating foliar symptoms. Root infection data will be needed to partition the resistance source in leaves, roots, or both.

QTL nomenclature

Another issue on SDS research related to mapping studies is the nomenclature of QTL. The Soybean Genetics Committee first allowed the Rfs (resistance to Fusarium solani f. sp. glycines) locus designation based on a single cross assay evaluated under greenhouse conditions. The Soybean Genetics Committee later allowed confirmed QTL to be re-named as cqSDS001 and cqSDS002. However, these nomenclatures were not consistently followed by later studies, and some literature reported the same QTL with different names, or the same names for different QTL. For example, Rfs1 and QRfs2 were both reported to be linked to Satt570 on chromosome (Chr) 18 (Abdelmajid et al. 2007; Lightfoot 2015; Triwitayakorn et al. 2005). The locus cqRfs4 (a.k.a. SDS 11-1 in SoyBase) overlaps with the locus QRfs5 (Abdelmajid et al. 2007; Kazi et al. 2008), while both qRfs5 and QRfs6 were reported to be linked to Satt354 on Chr20 (Abdelmajid et al. 2007; Lightfoot 2015; Luckew et al. 2013). Instead of locating near QRfs6, qRfs6 was reported to be linked to Satt549 on Chr03 (Lightfoot 2015), which is closer to Satt387 that links to QRfs7 (Abdelmajid et al. 2007).

In another example, QRfs7 was reported to either be located between Satt080 and Satt387 on Chr3 or located between Satt160 and Satt252 on Chr13 (Abdelmajid et al. 2007), while qRfs7 was reported to be linked to Satt226 and Sat_001 on Chr17 (Lightfoot 2015). QRfs8 was shown to be located in the interval of Satt160 and Satt252 on Chr13 (Abdelmajid et al. 2007), and qRfs12 was also reported to be linked to Satt160 (Lightfoot 2015; Luckew et al. 2013). Additional confusion derives from reports that Rfs12 was linked to Satt353 on Chr12, but cqRfs12 was mapped between Satt155 and Satt300 on Chr05 (Lightfoot 2015). Rfs16 was also reported to be linked to Satt353 on Chr12 without mentioning Rfs12 (Luckew et al. 2013). As for qRfs13 and cqRfs14, both were reported to be linked to Satt506 on Chr02 (Lightfoot 2015). These examples demonstrate the complicated nomenclature, the confusing map locations of QTL associated with resistance, and the inconsistencies of the QTL literature.

Additionally, some QTLs submitted to SoyBase were designated by a number referring to peer-reviewed publications and the order of submission to SoyBase. Under this scenario, SoyBase assigned QTL such as SDS6-1 or SDS18-2. The problem is that not all SDS mapping publications are included in SoyBase and the nomenclature does not provide the QTL location in the soybean genome. With more and more mapping studies, different names have been used in the literature according to the authors’ discretion instead of a systematic nomenclature (Abdelmajid et al. 2007; Anderson et al. 2015; Kassem et al. 2006; Luckew et al. 2017).

SDS disease resistance research: reorganization

Integration of SDS resistance loci in the soybean genome

The concerns described above provided the motivation for this review, which resulted in the proposal of an organized and simplified nomenclature for SDS resistance loci in the soybean genome based on specific rules. Typically, a meta-analysis is used to compile QTL from multiple studies and this approach has been used for different crops, such as barley, maize, rice and wheat (Ballini et al. 2008, Schweizer and Stein 2011; Soriano and Royo 2015; Wang et al. 2016). Most algorithms and tools of a meta-analysis were developed to assemble QTL from linkage maps (Goffinet and Gerber 2000; Veyrieras et al. 2007), and basic information such as the logarithm of the odds (LOD), the coefficient of determination (R2), and the positions of flanking markers of each QTL are required for methods such as BioMercator (Vasconcellos et al. 2017; Sosnowski et al. 2012). Unfortunately, not all SDS mapping studies presented these essential details for a meta-analysis, and there is no meta-analysis algorithm designed to combine QTL and SNPs for SDS resistance in soybean. Therefore, a manual assembly was more practical to visualize common regions repeatedly discovered for SDS resistance from linkage mapping and GWAS.

The manual assembly includes: (1) QTL with physical positions for both flanking markers. The intervals of such QTL are directly illustrated in the schematic ideogram; (2) QTL with known position for one flanking marker but unknown position for another flanking marker. The intervals of such QTL are illustrated using the known position and using the closest marker for the later case; and (3) SNPs or SSR markers were illustrated based on their locations. Our analysis compiles QTL and SNPs reported in the soybean SDS literature published till 2017, and generates a schematic ideogram using the physical map of soybean cultivar ‘Williams 82’ genome assembly 2 version 1 (Fig. 1; Table S1). A gray background is added to highlight each data entry, and a darker gray shading represents a region with multiple literature support.

Fig. 1
figure 1figure 1

Schematic ideogram for SDS resistance in soybean genome. Each gray line or block represents an SNP/marker or a QTL interval. Darker background indicates more SNPs, markers, or QTL were reported in the region. However, only regions repeatedly discovered by three different literature citations were considered as an Rfv (resistance to Fusarium virguliforme) locus, and are highlighted within red blocks

Full size image

The genomic picture provides an intuitive visualization of regions that have been repeatedly mapped and reported in the literature. There are ten regions reported by three or more studies (Table 1). We propose to name these regions as Rfv (resistance to F. virguliforme), followed by the Chr number. If two regions located on one Chr contain SDS resistance data, the region from the earliest publication receives the suffix priority. If two studies are published in the same year (e.g., two regions of Rfv18), the region with physical position closer to the beginning of the chromosome will receive the suffix priority (e.g., Rfv18-01 and Rfv18-02).

Table 1 Rfv (resistance to Fusarium virguliforme) loci for sudden death syndrome resistance
Full size table

However, there are two exceptions. One is a region on Chr13 close to qRfs8 and qRfs12 and the other is a region on Chr16 close to qRfs10. These two regions were not assigned an Rfv name because the third publication supporting these regions provides only one SSR marker instead of an interval. Under this situation, naming these two cases will result in an Rfv locus with a size of the third overlapping SSR marker length (89 and 56bp, respectively) (Table S1). As the objective of this review in proposing the Rfv nomenclature is to simplify the current system and offer an intuitive understanding, the inconclusive mapping results, therefore, remain recorded as is until future studies provide additional evidence for them.

Pleiotropic resistance or confounded mapping for Rfv18-01 and rhg1

Rfv18-01 (2.6Mb, overlapping with previously reported loci SDS 2-1, SDS 2-2, SDS 3-1, SDS 3-2, SDS 4-1, SDS 4-3, SDS 5-1, SDS 6-2, SDS 7-1, SDS 7-2, SDS 7-3, SDS 8-1, Rfs1, qRfs1, cqRfs1, Rfs2, and qRfs2) received the most literature citations or support (Table 1). In this locus, most studies phenotyped SDS foliar symptoms in fields with historical SDS incidence (Chang et al. 1996; Hnetkovsky et al. 1996; Iqbal et al. 2001; Abdelmajid et al. 2007; Kazi et al. 2008; Lightfoot 2015; Meksem et al. 1999; Njiti et al. 1998, 2002; Prabhu et al. 1999; Wen et al. 2014), except for one greenhouse study that used inoculated sand soil in addition to field trial data (Yuan et al. 2012).

Because multiple evidence supports the disease synergism between SCN and SDS (Chang et al. 1997; Brzostowski et al. 2014; Gao et al. 2006; Kandel et al. 2017; Roth et al. in press; Westphal et al. 2014), it raises the question whether SDS foliar symptoms in field studies were biased by SCN-infested fields, especially when a major SCN resistance source, such as the rhg1 locus (Caldwell et al. 1960), was detected in the same region as Rfv18-01. Therefore, it is possible that SCN susceptible soybean genotypes display more foliar symptoms due to higher SDS pressure resulting from co-infection with SCN. Most linkage mapping results supporting the Rfv18-01 were based on three bi-parental populations, ‘Forrest’ × ‘Essex’, ‘Hartwig’ × ‘Flyer’, and ‘Pyramid’ × ‘Douglas’ (Table 1). The parents ‘Forrest’, ‘Hartwig’, and ‘Pyramid’ are resistant to both SCN and SDS (Kazi et al. 2010; Meksem et al. 2001; Njiti et al. 1996), which do not contribute to the clarification for which susceptibility is being evaluated if the population was screening in fields with the presence of both diseases. The USDA GRIN database indicates the pedigree of ‘Forrest’ (PI 548655) has the ‘Peking’ (PI 438497) SCN resistance source. Pedigree information of both ‘Hartwig’ (PI 543795) and ‘Pyramid’ (PI 512039) has ‘Forrest’ in the genetic background. ‘Pyramid’ also has the PI 88788 SCN resistance source in its pedigree (Fig. S1). It is known that SCN resistance in PI 88788 has the rhg1-b allele, and SCN resistance in ‘Peking’ has both rhg1-a and Rhg4 (Liu et al. 2017). Therefore, genetic segregation of these progeny populations used in SDS mapping all carry segregation differences for SCN resistance. The dual resistance segregation for SCN and SDS could interfere with phenotyping, leading to the same resistance locus if the SDS disease evaluations were also performed on SCN-infested fields.

Current understanding of SCN resistance from rhg1-b lies on copy number variation (Cook et al. 2012), which differs from conventional understanding of disease resistance that commonly relies on leucine-repeat rich (LRR)-coding genes. There are several LRR-coding genes close to rhg1 (Chang et al. 2016a), and this region was also reported to control pleiotropic resistance for rhg1-a/Rfs2 resistance to SCN and SDS, respectively (Srour et al. 2012). Two putative soybean genes including an LRR-coding receptor like kinase gene (Glyma.18g023500) and a laccase (Glyma.18g023600) were reported to have dual partial resistance to SCN and SDS (Afzal et al. 2013; Iqbal et al. 2008). Moreover, a recent study pointed out that rhg1-a interacts with another SCN-resistant locus Rhg4 located on Chr08 (Liu et al. 2017), and the region (Chr08, approx. 8.07–10.77Mb) has been mapped for SDS in two studies (Prabhu et al. 1999; Swaminathan et al. 2016). Although SDS field phenotyping raises the concern with SCN interference, the possibility of pleiotropic resistance to both SCN and SDS requires additional research for clarification.

Most of the proposed Rfv loci (Table 1) were identified based on disease assessment using field trials; some Rfv loci were phenotyped using the stem cutting assay with culture filtrates or F. virguliforme phytotoxins (Chang et al. 2016b; Swaminathan et al. 2016). Rfv06-01 (6.92Mb, overlapping with previously reported loci SDS 1-1, SDS 1-3, SDS 2-5, SDS 2-6, SDS 7-5, SDS 8-2, qDX006, qDX007, qDX008, Rfs4 and qRfs4) is one locus supported by a stem cutting assay (Swaminathan et al. 2016) and several field studies (Abdelmajid et al. 2007; Anderson et al. 2015; Chang et al. 1996; Hnetkovsky et al. 1996; Iqbal et al. 2001; Kassem et al. 2006; Lightfoot 2015; Njiti et al. 2002; Wen et al. 2014). Rfv06-02 (7.53Mb, overlapping with previously reported loci SDS 11-1, SDS 16-5, SDS 16-6, ds1, cqRfs4, and qRfs5), Rfv09-01 (0.98Mb, overlapping with previously reported loci SDS 16-1, qDX009, and qRfs13), Rfv13-01 (2.83Mb, overlapping with previously reported loci SDS 15-1, SDS 16-8, Dl1, qRfs9, sdsyld2, sdsyld3), and Rfv16-01 (0.51Mb, overlapping with previously reported loci SDS 15-7, qDX007, and an unnamed locus) were also found in the same stem cutting study (Swaminathan et al. 2016), and by field studies (Abdelmajid et al. 2007; Anderson et al. 2015; Kassem et al. 2006; Kazi et al. 2008; Yuan et al. 2012). The overlapping results highlight the possibility that these five loci are regions that harbor foliar resistance to phytotoxins.

While many LRR-coding genes and disease resistance genes are located within the Rfv06-01 locus, the region also contains an aldo–keto reductase gene (Glyma.06g260200) that displays differential expression in response to phytotoxin treatments (Radwan et al. 2013). The Rfv06-02 locus contains a large interval with many genes including LRR-coding genes, but the only genes with differential expression in response to phytotoxins were a homeobox transcription factor (Glyma.06g190200), and a zinc finger transcription factor (Glyma.06g196600) (Radwan et al. 2013). Other mapping studies also pointed out regions on Chr06 for phytotoxicity and SDS foliar responses (Abdelmajid et al. 2012; Bao et al. 2015; Chang et al. 2016b; Wen et al. 2014).

The Rfv09-01 region contains F-box domain genes but not LRR-coding genes, and no differentially expressed genes have been found in this region (Radwan et al. 2013). The interval of Rfv13-01 contains one differentially expressed gene encoding a carotenoid dioxygenase (Glyma.13g202200) (Radwan et al. 2013). The Rfv13-01 region is a gene hot zone with 26 LRR-coding genes mostly located within 29.79–30.92MB. There are two LRR-coding genes (Glyma.16g059200 and Glyma.16g059700) in the region of Rfv16-01, but similarly to the Rfv09-01 region, no differentially expressed genes were observed in it (Radwan et al. 2013). More research is needed to confirm if any LRR-coding genes, differentially expressed genes, or other genes within these loci might be the source of foliar resistance to phytotoxicity.

Conclusion

In this review, ten loci were assigned to the Rfv nomenclature on the basis that the loci were supported by three or more publications. Although this proposed nomenclature does not provide a gene validation per se, the criteria on the basis of number of publications reporting on the same QTL provide a higher level of confidence regarding QTL mapping reproducibility. SDS is a continued threat to soybean production causing substantial yield reductions and economic losses of up to $669.2 million dollars (Allen et al. 2017; Koenning and Wrather 2010; Navi and Yang 2016). As such we prepared this review on the current status of SDS quantitative resistance knowledge and proposed a unified SDS resistance loci (Rfv) nomenclature for breeders and pathologists working to improve SDS resistance.

Author contribution statement

HXC integrated the literature, presented the data, drafted and wrote the manuscript. MGR reviewed the data and contributed to the manuscript. DW, SRC, DAL, GLH, and MIC reviewed and contributed to the manuscript.